Trickiness with unittesting

[arockstar at gmail.com]

Basically, I'm trying to implement a setUp() and TearDown() for a
python TestSuite (as opposed to an individual test within the suite).

Sort of.

I have a few different test suites (call them SuiteA, SuiteB,...). For
one of the test suites (SuiteA), I need to execute a bit of code (say
startFoo()) before the first test in SuiteA runs, and once when the
last test of SuiteA finishes (endFoo()).

Making this even trickier is that the testing framework combines the
different suites (SuiteA, SuiteB) as well as individual tests into
master test suite which my testrunner then runs.

Also complicating matters is the fact that the testing framework can
run an *individual* test from any TestSuite: in that case, if I'm
running SuiteA.test1, I'd want to call startFoo() and endFoo() as
well.

I could use a global variable as a flag and see if it's been set
during the setUp() method of the test cases (which would allow me to
call startFoo() once when tests from SuiteA are run from the master
suite), but that doesn't help me with calling endFoo() when the tests
from SuiteA finish. Plus, it's not very elegant.

So:
Given a master test suite comprised of a bunch of different tests from
different suites, I need to call startFoo() before tests from SuiteA
are run, and stopFoo() after all the tests from SuiteA have finished
running.